The present invention relates generally to web Application Programming Interface (API) recommendations, and more particularly, but not by way of limitation, to a system, a method, and a recording medium for recommending a web API using consumer-provided content.
Conventional systems provide matching between quality of service (QoS) information exposed by providers and required by consumers respectively, or the descriptions of services created by service providers. However, the conventional systems result in information asymmetry in that providers aim to convince consumers of their offerings, and are thus likely to expose biased information. Furthermore, the conventional systems' information is static over time.
Other conventional systems merely characterize the latent topics between service descriptions and queries, and then propose a matching to the user's needs based on the topic relevance. However, the conventional systems directly use the provider's web API descriptions which conventionally have not led the user to the correct web API for their needs.
That is, the above conventional systems, and other conventional web API recommendation systems are limited in their application in that they are only directed to viewing data provided by the web API providers and attempt to match a user with an appropriate web API based on the providers descriptions.
Thus, there is a technical problem in the conventional systems that the methods of matching a user with an appropriate web API account, specific to the user's requirements using the provider side information, can be biased, static, and use specific market tested wording to convince users to consume the specific web API when the web API may not be the best match for the user.